Neuronal heterogeneity of normalization strength in a circuit model

Area of Science:

• Computational neuroscience
• Systems neuroscience
• Visual cortex function

Background:

• Neurons in higher-order visual areas use normalization for information integration.
• Normalization strength varies significantly across neurons.
• This heterogeneity correlates with attention-modulated neural responses, but its circuit basis is unknown.

Purpose of the Study:

• To investigate the circuit mechanisms underlying heterogeneous normalization strength in the visual cortex.
• To explore the computational consequences of this heterogeneity.

Main Methods:

• Developed a spiking neuron network model of the visual cortex.
• Analyzed the relationship between inhibitory currents and normalization strength.
• Examined information encoding efficiency and network capacity.

Main Results:

• Normalization strength heterogeneity strongly correlates with the inhibitory current received by neurons.
• This correlation explains observed relationships between normalization and spike count correlations.
• Neurons with stronger normalization exhibit more efficient information encoding.

Conclusions:

• The model provides a mechanistic explanation for heterogeneous normalization in the visual cortex.
• Neuronal heterogeneity in normalization strength enhances network-level information processing and capacity.
• This highlights the computational benefits of neuronal diversity in sensory processing.